[llvm-testresults] OldGrawp-NO-LTO-PIC i386 nightly tester results

Apache apache at cs.uiuc.edu
Sat Feb 16 05:26:57 PST 2008


http://llvm.org/nightlytest/test.php?machine=233&night=4966
Name: il0102a-dhcp80.apple.com
Nickname: OldGrawp-NO-LTO-PIC
Buildstatus: OK

New Test Passes:
None

New Test Failures:
test/C++Frontend/2008-02-13-sret.cpp [DEJAGNU]


Added Tests:
test/Assembler/2004-03-07-FunctionAddressAlignment.ll
test/CodeGen/Alpha/mb.ll
test/CodeGen/Alpha/wmb.ll
test/CodeGen/X86/lfence.ll
test/CodeGen/X86/mfence.ll
test/CodeGen/X86/nofence.ll
test/CodeGen/X86/sfence.ll


Removed Tests:
test/Assembler/2004-03-07-FunctionAddressAlignment.llx
test/CodeGen/X86/coalescer-commute2.ll


Significant changes in test results:
JIT:
 singlesource/Benchmarks/Shootout-C++/methcall: -16.88% (10.90 => 12.74)
 external/SPEC/CFP2006/433.milc/433.milc: -7.84% (34.05 => 36.72)
LLC compile:
 multisource/Applications/kimwitu++/kc: 11.93% (11.6764 => 10.2830)
 multisource/Benchmarks/MallocBench/gs/gs: 8.36% (2.9664 => 2.7185)
 external/SPEC/CFP2000/177.mesa/177.mesa: 5.52% (6.6895 => 6.3200)
 external/SPEC/CFP2006/447.dealII/447.dealII: 7.34% (48.6641 => 45.0929)
 external/SPEC/CINT2000/252.eon/252.eon: 9.09% (4.1505 => 3.7734)
 external/SPEC/CINT2000/255.vortex/255.vortex: 5.37% (6.6745 => 6.3163)
LLC-BETA compile:
 multisource/Applications/kimwitu++/kc: 11.93% (11.7019 => 10.3061)
 multisource/Benchmarks/MallocBench/gs/gs: 8.14% (2.9647 => 2.7234)
 external/SPEC/CFP2000/177.mesa/177.mesa: 5.74% (6.7142 => 6.3287)
 external/SPEC/CINT2000/252.eon/252.eon: 8.86% (4.1561 => 3.7879)
 external/SPEC/CINT2000/255.vortex/255.vortex: 5.46% (6.6844 => 6.3197)
 external/SPEC/CINT2006/400.perlbench/400.perlbench: 5.04% (21.0968 => 20.0329)
LLC:
 multisource/Benchmarks/ASCI_Purple/SMG2000/smg2000: -5.97% (9.05 => 9.59)
 external/SPEC/CFP2006/433.milc/433.milc: -7.23% (33.74 => 36.18)
CBE:
 external/SPEC/CFP2006/433.milc/433.milc: -5.95% (33.11 => 35.08)
LLC-BETA:
 external/SPEC/CFP2006/433.milc/433.milc: -8.09% (33.76 => 36.49)




More information about the llvm-testresults mailing list